Minimal residual methods for solving a class of R-linear systems of equations
Kui Du, Olavi Nevanlinna
Abstract
Kui Du, Olavi Nevanlinna
Abstract
Recently, R-linear GMRES was proposed by Eirola, Huhtanen and von Pfaler [SIAM J. Matrix Anal. Appl. 25 (2004), pp. 804–828] for solving a class of R-linear systems of equations. In this work we investigate R-linear GMRES through the equivalent real formulations of the R-linear system. We show that R-linear GMRES requires fewer matrix-vector products than GMRES applied to the related C-linear system. Numerical results for an artificial example and the inverse problem of reconstructing an unknown electric conductivity are reported to confirm our theoretical results. AMS subject classifications: 65F10, 15A18
OpenAlex reports 2 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Recently, R-linear GMRES was proposed by Eirola, Huhtanen and von Pfaler [SIAM J. Matrix Anal. Appl. 25 (2004), pp. 804–828] for solving a class of R-linear systems of equations. In this work we investigate R-linear GMRES through the equivalent real formulations of the R-linear system. We show that R-linear GMRES requires fewer matrix-vector products than GMRES applied to the related C-linear system. Numerical results for an artificial example and the inverse problem of reconstructing an unknown electric conductivity are reported to confirm our theoretical results. AMS subject classifications: 65F10, 15A18
Key concepts: Generalized minimal residual method, Linear system, Mathematics, System of linear equations, Coefficient matrix, Applied mathematics, Linear equation, Matrix (chemical analysis)